Combined transformations are easier to manage when their effect is written as a mapping of points. A horizontal reflection, scaling and translation can then be handled together, without relying on a memorised order that may not match the algebraic form.

This lesson extends Transformations of Functions to mapped domains and ranges, absolute-value folding, asymptotes and restricted logarithmic or radical inputs.

On this page

Derive the complete point map

Let

g(x)=Af(B(x−h))+k,A≠0,B≠0.g(x)=A f\bigl(B(x-h)\bigr)+k,\qquad A\ne0,\ B\ne0.

Suppose (u,v)(u,v) lies on the graph of ff, so v=f(u)v=f(u). To use the same original input uu inside gg, solve B(x−h)=uB(x-h)=u. Then

x=h+uB,g(x)=Av+k.x=h+\frac uB,\qquad g(x)=Av+k.

The complete point mapping is therefore

(u,v)↦(h+uB,k+Av).(u,v)\longmapsto\left(h+\frac uB,\ k+Av\right).

Horizontal distances scale by 1/|B|1/|B|, with a reversal when B<0B<0. Vertical distances scale by |A||A|, with a reversal when A<0A<0. The parameters h,kh,k locate the translations in this factored form.

If A=0A=0, the output is constant on the original composite’s domain, not necessarily on every real input. If B=0B=0, the inner expression is constant and must belong to the domain of ff. The ordinary invertible point map does not apply to those degenerate cases.

Map a radical graph with two reversals

Take f(u)=uf(u)=\sqrt u and

g(x)=−2−3(x−1)+4.g(x)=-2\sqrt{-3(x-1)}+4.

Here A=−2A=-2, B=−3B=-3, h=1h=1, k=4k=4, giving

(u,v)↦(1−u/3,4−2v).(u,v)\longmapsto(1-u/3,4-2v).

The parent points (0,0),(1,1),(4,2),(9,3)(0,0),(1,1),(4,2),(9,3) become

(1,4),(2/3,2),(−1/3,0),(−2,−2).(1,4),\quad(2/3,2),\quad(-1/3,0),\quad(-2,-2).

Each image point can be checked in the displayed formula. The domain comes from −3(x−1)≥0-3(x-1)\ge0, so x≤1x\le1. Since the square root is nonnegative, g(x)≤4g(x)\le4. Every value at or below 4 is attained: for a desired output y≤4y\le4, set −3(x−1)=(4−y)/2\sqrt{-3(x-1)}=(4-y)/2 and solve for xx.

Transform the original input and output separately

Combined transformations map parent points exactlyThe square root parent and its transformed graph minus two times the square root of minus three times x-1 plus four are shown. The transformed endpoint is (1,4), and parent input nine maps to x=-2.f(u) = √ug(x) = −2√(−3(x−1)) + 4-2024-20246810OutputInput(1, 4)(−2, −2)
The two coordinate axes use equal physical units. Dashed and solid lines distinguish the graphs without colour.
Diagram explanation and exact data

Parent point (u,v) maps to (1-u/3,4-2v). The original domain u≥0 maps to x≤1, and the range v≥0 maps to y≤4.

Transform the original input and output separately: exact reference data
Parent point Image point
(0,0)(0,0) (1,4)(1,4)
(1,1)(1,1) (2/3,2)(2/3,2)
(4,2)(4,2) (−1/3,0)(-1/3,0)
(9,3)(9,3) (−2,−2)(-2,-2)

The graph’s endpoint is (1,4)(1,4), not (3,4)(3,4) or (1,−4)(1,-4). Working with the point map prevents the two negative factors from being confused.

Transform intervals with their order and endpoints

Suppose a parent function is restricted to domain [−3,6][-3,6] and has range [0,5][0,5]. Under the same mapping,

x=1−u/3,y=4−2v.x=1-u/3,\qquad y=4-2v.

The input endpoints u=−3,6u=-3,6 map to x=2,−1x=2,-1. Reorder them when writing the interval: the transformed domain is [−1,2][-1,2]. The output endpoints 0 and 5 map to 4 and −6, so the range is [−6,4][-6,4].

Negative scale factors reverse endpoint order but do not change whether a given endpoint is included. If the original domain were (−3,6](-3,6], its image would be [−1,2)[-1,2): the image of the excluded endpoint remains excluded.

This example concerns a specified bounded parent domain and range. It is a different parent-function situation from the unrestricted square-root example above. Domain and range information must accompany the chosen function, not be borrowed from a neighbouring example.

Factor the inner expression before reading a shift

For

g(x)=2f(3x+6)−1,g(x)=2f(3x+6)-1,

factor the input as 3(x+2)3(x+2). In the standard form, B=3B=3 and h=−2h=-2, giving the map

(u,v)↦(−2+u3,2v−1).(u,v)\longmapsto\left(-2+\frac u3,2v-1\right).

The horizontal translation is −2, not +6. One valid construction compresses the parent graph horizontally by a factor of 1/31/3, then shifts left two units. Shifting first and then compressing by the same stated amounts would usually give a different function.

The algebra determines the correct mapping. A verbal list of transformations is useful only when its order produces that mapping. OpenStax’s transformation chapter provides the basic transformations being combined here.

Distinguish folding outputs from folding inputs

For g(x)=|f(x)|g(x)=|f(x)|, keep each input and replace its output by its absolute value. Portions below the x-axis reflect upward; portions on or above it stay unchanged. Thus

|x2−4|={x2−4,|x|≥2,4−x2,|x|<2.|x^2-4|=\begin{cases}x^2-4,&|x|\ge2,\\4-x^2,&|x|<2.\end{cases}

By contrast, g(x)=f(|x|)g(x)=f(|x|) uses the nonnegative side of the parent function as the source and mirrors it across the y-axis. For f(x)=x−1f(x)=x-1,

f(|x|)=|x|−1,f(|x|)=|x|-1,

whereas |f(x)|=|x−1||f(x)|=|x-1|. Their vertices are (0,−1)(0,-1) and (1,0)(1,0) respectively, so they are plainly different graphs.

The domain of f(|x|)f(|x|) is {x:|x|∈domf}\{x:|x|\in\operatorname{dom}f\}. If the parent exists only for negative inputs, folding the input might leave no allowed input. Do not assume every absolute-value transformation automatically has domain ℝ\mathbb R.

Map asymptotes and preserve forbidden inputs

Start with f(u)=1/uf(u)=1/u and define

g(x)=32x−4+1=3f(2(x−2))+1.g(x)=\frac3{2x-4}+1=3f\bigl(2(x-2)\bigr)+1.

The input u=0u=0 maps to x=2x=2, so the vertical asymptote is x=2x=2. The horizontal asymptote v=0v=0 maps to y=1y=1. Directly,

g(x)−1=32x−4→0g(x)-1=\frac3{2x-4}\longrightarrow0

at both ends. The point map is (u,v)↦(2+u/2,1+3v)(u,v)\mapsto(2+u/2,1+3v).

If the parent had an additional excluded point, that exclusion would map too. A simplified transformed formula should not silently restore it. Asymptotes describe limiting behaviour; holes describe missing points with finite nearby limits.

Check logarithms and exponentials after transformation

The transformed logarithm ln(4−2x)\ln(4-2x) requires 4−2x>04-2x>0, giving x<2x<2. Its vertical boundary is excluded. This strict condition differs from the inclusive endpoint of a square root, whose argument may equal zero.

For

q(x)=−2⋅3−(x+1)+5,q(x)=-2\cdot3^{-(x+1)}+5,

the domain is all real numbers and the range is (−∞,5)(-\infty,5). As xx increases, 3−(x+1)3^{-(x+1)} decreases, and multiplication by −2 reverses that trend. The final function is increasing, although both a horizontal and vertical reflection occur in its construction.

The output approaches 5 from below as x→+∞x\to+\infty. Setting q(x)=0q(x)=0 gives 3−(x+1)=5/23^{-(x+1)}=5/2, hence x=−1−ln(5/2)/ln3x=-1-\ln(5/2)/\ln3. A graph label should be placed using this actual intercept, not an approximate visual guess.

Practice questions

  1. A parent point is (6,−1)(6,-1). Find its image under g(x)=−2f(−3(x−1))+4g(x)=-2f(-3(x-1))+4.
  2. Under the same transformation, what happens to the parent domain [−3,6][-3,6] and range [0,5][0,5]?
  3. Find the real domain of ln(4−2x)\ln(4-2x).
Check your answers
  1. The map gives (1−6/3,4−2(−1))=(−1,6)(1-6/3,4-2(-1))=(-1,6). The original input and output transform by different formulas.
  2. The domain endpoints map to 2 and −1, producing [−1,2][-1,2]. The range endpoints map to 4 and −6, producing [−6,4][-6,4]. All endpoints remain included.
  3. Require 4−2x>04-2x>0, so x<2x<2. Equality would give the logarithm of zero and is not permitted.

Related topics

Composite and Inverse Functions: Advanced Problems reverses and combines these mappings while retaining the correct domains. Asymptotes develops the limiting behaviour of transformed rational functions.

Sources and further reading